To be clear, by 'infrastructure' I mean physical lines going to the premises. There's still hope for wireless, but competition in the physical connection space is effectively dead, and getting deader with every merger.
But also: why are the physical lines a separate market from wireless? They are very, very close substitutes – bits are bits. And advancing tech or clever to-the-curb (but not to-the-premises) buildouts could make wireless broadband fully competitive with the wired networks.
If the wired incumbents get fat and happy on a position of temporary advantage – for example for bulk video delivery – then those potential wireless advances will see more investment.
These are different market. You may think the buck stops somewhere, but after 720p video streaming, you get 1080. Then you get gaming, at sixty FPS, not 30. Then you get VR streaming, up to 10 times the pixel rate, and nearly twice the frame rate to have perfect presence. And never mind the latency requirements. Then you can get more sci-fi, with direct neural interfaces, and, eventually whole brain emulation over the network —though at that point, the economy would have changed so much that the fibre vs wireless debate will be quite moot.
That's effectively what I'm advocating--by removing their monopoly power to be the only ISP available to the customer on the other end of their lines. Breaking them up into regionalized monopolies doesn't give consumers a choice, but opening the lines to competition does.
And if it doesn't? How long do we wait?
For example, what if ISPs roll out traffic-preferences-for-pay... but the base bandwidth available to customers, and sites that don't opt to pay for preference, also keeps improving, on a price/performance basis? Would that still be an emergency requiring novel nationwide regulation?
And wouldn't we learn something if some local markets showed abuse or lagging price/performance, while others thrived? Couldn't we, and especially the lagging jurisdictions, then do the minimal targeted reforms to improve results based on actual history, rather than a theoretical "neutral net" ideal?
Note that unlike the shared broadcast airwaves, the economic case for wires being a 'commons' needing public management is weak. There's already at least 3 sets of wires running to most structures – electricity, phones, and cable – and more could be added if those aren't satisfying the customers. And even electricity – far more foundational than internet – is regulated in a diverse manner locally, not nationally.
Wait and see and let localities and businesses experiment is just fine - it's what got us to this point, after all, and the internet is doing just fine – albeit prone to occasional panic attacks: "Death of the Internet Predicted, Film At 11".
Specific measurable abuses and the court rulings in the specific attempt to deal with them in the Comcast FCC action and subsequent litigation in 2007-2010 was one of the things which spurred the FCC to move network neutrality from a set of generalized objectives that it was seeking to preserve in it its other actions to a specific focus of dedicated, generally-applicable regulation.
So been there, done that.
So I'm not sure this is a 'natural monopoly' at all. There's been too little time to observe, and things are still changing fast.
Let abusive incumbents try to collect monopoly rents. That would attract lots of new solutions from very savvy and deep-pocketed challengers, like Google, Apple, and other strategic investors.